Did any subgroup show harm from semaglutide in SUSTAIN-6?

No. Every pre-specified subgroup had a point estimate for MACE below 1.0, meaning all trended toward benefit. No confidence interval suggested increased cardiovascular risk in any demographic or clinical stratum.

Was the MACE reduction with semaglutide different in men versus women?

The point estimates were similar (HR ~0.75 in men, ~0.73 in women), and the interaction p-value was not significant. The wider confidence interval in women reflects a smaller sample (39% female), not a reduced treatment effect.

Do older patients benefit more from semaglutide than younger patients?

The point estimate was slightly lower (HR ~0.69) in patients aged 65 and older compared with those under 65 (HR ~0.82), but the interaction test was not significant (p = 0.52). Older patients likely benefit at least as much, and possibly more in absolute terms because of higher baseline risk.

Does BMI affect semaglutide's cardiovascular benefit?

Patients with BMI ≥ 30 showed a numerically stronger HR (~0.70) than those below 30 (~0.84). The interaction was not significant. Weight loss may contribute to the mechanism, but the trial was not designed to isolate weight-mediated cardioprotection from glycemic or anti-inflammatory effects.

Can patients with CKD stage 3 expect cardiovascular benefit from semaglutide?

SUSTAIN-6 subgroup data showed a favorable point estimate (HR ~0.71) for patients with eGFR 30, 59, though the confidence interval crossed 1.0. The FLOW trial later confirmed renal and cardiovascular benefits of semaglutide in CKD patients with T2D.

Why didn't SUSTAIN-6 report race-specific MACE outcomes in detail?

The cohort was 83% White, limiting the statistical power for race-stratified analyses. Published forest plots did not provide the same granularity for race as for age, sex, or BMI. Larger, more diverse trials like SOUL are expected to address this gap.

Is the 0.5 mg dose of semaglutide enough for cardiovascular protection?

Both 0.5 mg and 1.0 mg doses showed MACE benefit in SUSTAIN-6, with HRs of approximately 0.77 and 0.74, respectively. The trial was not powered to detect a dose-response difference, but both doses provided a clear signal.

How does the SUSTAIN-6 subgroup picture compare with SELECT?

SELECT (2023) tested semaglutide 2.4 mg in 17,604 patients with obesity but without diabetes. Its larger size and longer follow-up produced tighter subgroup confidence intervals. The consistency of benefit across subgroups in both trials strengthens the case that semaglutide's CV effect is broadly applicable.

Should SUSTAIN-6 change prescribing for primary prevention of cardiovascular disease?

Not directly. Only 17% of participants lacked established CVD, and the subgroup HR (0.87) had a wide confidence interval. The FDA's cardiovascular indication for Ozempic is limited to patients with established CVD. Primary prevention remains an evidence gap for this dose and formulation.

Were the subgroup analyses in SUSTAIN-6 adjusted for multiple comparisons?

No. Following standard CVOT methodology, interaction tests were performed without multiplicity adjustment. All subgroup results should be considered exploratory and hypothesis-generating rather than confirmatory.

SUSTAIN-6 Subgroup Analyses: Who Responded Most and Least

By HealthRX Medical Team

Published May 25, 2026Updated May 25, 2026Last reviewed May 25, 2026

At a glance

| Parameter | Detail | |---|---| | N | 3,297 | | Intervention | Semaglutide 0.5 mg or 1.0 mg SC once weekly | | Comparator | Placebo (volume-matched) | | Duration | Median 2.1 years | | Primary endpoint | First occurrence of 3-point MACE (CV death, nonfatal MI, nonfatal stroke) | | Key result | HR 0.74 (95% CI 0.58, 0.95; p < 0.001 for non-inferiority) | | Trial design | Randomized, double-blind, placebo-controlled CVOT |

Why Subgroup Analyses Matter Here

SUSTAIN-6 was designed as a cardiovascular outcomes trial (CVOT) required by the FDA for GLP-1 receptor agonist approval. The overall 26% relative risk reduction in 3-point MACE established semaglutide's cardiovascular credentials. But the overall hazard ratio tells a prescriber nothing about the 62-year-old woman with an eGFR of 48 sitting in the exam room versus the 54-year-old man with a BMI of 44 and no prior MI.

Pre-specified subgroup analyses were built into the SUSTAIN-6 statistical analysis plan precisely to answer that kind of question. Post-hoc explorations published subsequently added granularity on stroke subtypes, retinopathy risk stratification, and renal endpoints. This page walks through both layers.

Trial Population at Baseline

Before interpreting any subgroup split, the denominator matters. SUSTAIN-6 enrolled adults with type 2 diabetes and HbA1c ≥ 7.0% who were either 50 years or older with established cardiovascular disease (83% of the cohort) or 60 years or older with at least one cardiovascular risk factor. This skewed the population toward high-risk, older patients with long disease duration (mean ~13.9 years of T2D). Roughly 61% were male, mean BMI was ~32.8 kg/m², and mean baseline HbA1c was ~8.7%.

The racial composition was predominantly White (~83%), with smaller representation from Asian (~8%), Black (~6%), and other groups. This demographic tilt limits the strength of race-stratified conclusions, a point the original publication acknowledged.

The HealthRX Subgroup Response Framework

To organize the pre-specified and post-hoc results, we scored each subgroup split on three axes: (1) the direction and magnitude of the point estimate relative to the overall HR of 0.74, (2) whether the 95% CI excluded 1.0 within that subgroup, and (3) the p-value for interaction (testing whether the treatment effect genuinely differed between subgroups versus reflecting random variation). A non-significant interaction p-value (typically > 0.10) means the subgroup difference is consistent with chance.

| Subgroup | Semaglutide HR (95% CI) | Interaction p | Signal | |---|---|---|---| | Age ≥ 65 y | ~0.69 (0.50, 0.95) | 0.52 | Consistent, slightly favored | | Age < 65 y | ~0.82 (0.56, 1.20) |, | Consistent, wider CI | | Male | ~0.75 (0.55, 1.01) | 0.78 | Consistent | | Female | ~0.73 (0.49, 1.10) |, | Consistent, underpowered | | BMI ≥ 30 kg/m² | ~0.70 (0.52, 0.94) | 0.46 | Consistent, slightly favored | | BMI < 30 kg/m² | ~0.84 (0.56, 1.25) |, | Consistent, wider CI | | Established CVD | ~0.72 (0.55, 0.93) | 0.61 | Consistent | | CV risk factors only | ~0.87 (0.48, 1.56) |, | Underpowered (17% of N) | | Baseline HbA1c ≥ 8.5% | ~0.70 (0.50, 0.98) | 0.55 | Consistent, slightly favored | | Baseline HbA1c < 8.5% | ~0.79 (0.55, 1.13) |, | Consistent | | eGFR 30, 59 (CKD 3) | ~0.71 (0.47, 1.08) | 0.72 | Consistent, underpowered | | eGFR ≥ 60 | ~0.76 (0.56, 1.02) |, | Consistent |

Note: exact subgroup CIs were reported in the supplementary appendix of the original NEJM publication. Values above reflect the forest plot data; minor rounding applies.

The critical finding: no interaction p-value reached significance. The treatment effect did not reliably differ across any tested split.

Who Responded Most (by Point Estimate)

Three subgroups showed numerically stronger point estimates:

Patients aged 65 and older. The HR dropped to approximately 0.69, slightly better than the overall 0.74. This is clinically logical. Older patients with T2D carry higher absolute MACE risk, so even a similar relative risk reduction translates to more events prevented. The 2019 ADA Standards of Care subsequently cited GLP-1 RA benefit in older adults with established CVD as a reason to prioritize this drug class in that population.

Patients with BMI ≥ 30 kg/m². The point estimate of ~0.70 slightly outperformed the leaner subgroup. Semaglutide's weight loss effect (mean −3.6 kg to −4.9 kg versus placebo in SUSTAIN-6) may contribute to cardioprotection beyond glycemic control. The SELECT trial, published later, would go on to confirm semaglutide's MACE benefit in patients with obesity and no diabetes, reinforcing that weight-mediated pathways matter.

Patients with higher baseline HbA1c (≥ 8.5%). Greater glycemic improvement leaves more room for plaque-stabilizing, anti-inflammatory, and hemodynamic benefits to accrue. The ~0.70 HR in this stratum is consistent with the hypothesis, though the interaction test did not confirm a true difference from the lower-HbA1c group.

Who Responded Least (by Point Estimate)

Patients without established CVD (risk factors only). The HR of ~0.87 with a wide confidence interval (0.48, 1.56) simply reflects statistical noise from a small cell. Only ~17% of participants fell into this category. This is not evidence of no benefit; it is evidence of insufficient power. The FDA's 2020 label update for Ozempic limited the cardiovascular indication to patients with established CVD rather than primary prevention, partly because SUSTAIN-6 could not confirm benefit in the lower-risk group.

Patients with BMI < 30 kg/m². The HR of ~0.84 was nominally weaker, but the confidence interval fully overlapped the ≥ 30 subgroup. With roughly 40% of the cohort below the obesity threshold, this subgroup was reasonably sized. The result suggests benefit persists in overweight-but-not-obese patients, though the magnitude is less certain.

Female participants. The HR of ~0.73 was nearly identical to the overall estimate, but the confidence interval crossed 1.0 because only 39% of participants were women. This is a power problem common to cardiovascular trials, not a signal of sex-based treatment resistance. Post-hoc pooled analyses across the SUSTAIN program have not identified a meaningful sex-by-treatment interaction.

Race and Ethnicity: The Data Gap

SUSTAIN-6 enrolled ~83% White participants, ~8% Asian, and ~6% Black. The published forest plots did not break out race-specific MACE hazard ratios with the same precision as age or BMI subgroups. This is a real limitation. GLP-1 receptor agonists have shown variable weight-loss efficacy across racial groups in obesity trials (smaller mean weight loss in Black participants in STEP trials, for example). Whether that translates to differential cardiovascular protection remains unknown from SUSTAIN-6 data alone.

The SOUL trial (semaglutide, cardiovascular and renal outcomes, reported 2024) enrolled a more diverse population and may eventually provide race-stratified MACE data to fill this gap.

Dose-Level Subgroup: 0.5 mg Versus 1.0 mg

SUSTAIN-6 randomized patients to semaglutide 0.5 mg or 1.0 mg (or corresponding placebo volumes). The trial was not powered to detect a dose-response for MACE, but the numerical results were suggestive:

| Dose | MACE HR vs. placebo | |---|---| | Semaglutide 0.5 mg | ~0.77 | | Semaglutide 1.0 mg | ~0.74 |

Both doses showed benefit. The modest separation does not support withholding the higher dose for cardiovascular purposes, and the Ozempic prescribing information recommends titrating to 1.0 mg for glycemic control. Clinicians using 0.5 mg for tolerability can be reassured that cardiovascular signal was present at that dose.

Renal Function Subgroups

Patients with eGFR 30 to 59 mL/min/1.73 m² (CKD stage 3) represented a meaningful fraction of the cohort. Their point estimate (~0.71) was numerically favorable, though the confidence interval crossed 1.0 because of smaller sample size. This matters because CKD accelerates cardiovascular risk in T2D, and prescribers often hesitate to use injectable GLP-1 RAs in patients with renal impairment.

Semaglutide does not require dose adjustment for renal function down to eGFR 15, per the FDA label. The SUSTAIN-6 subgroup data, while underpowered, align with the pharmacokinetic rationale. The FLOW trial (2024) later demonstrated renal-specific benefit of semaglutide in CKD patients with T2D, reinforcing this signal.

Methodological Caveats

Pre-specified subgroup analyses in a trial powered for a composite endpoint carry inherent limitations:

Multiplicity. Testing a dozen subgroups inflates false-positive risk. SUSTAIN-6 did not adjust for multiplicity in subgroup analyses, following standard CVOT practice. The interaction p-values are exploratory.
Asymmetric cell sizes. The 83/17 split between established CVD and risk-factors-only groups made the smaller cell nearly uninterpretable.
Short median follow-up. At 2.1 years, some subgroups simply did not accumulate enough events to produce stable estimates. The SELECT trial (mean follow-up 39.8 months, N = 17,604) later showed how longer observation and larger sample size tightened subgroup confidence intervals.
Post-hoc analyses were not pre-registered. Stroke subtype and retinopathy-by-baseline-status analyses published after the primary paper should be interpreted as hypothesis-generating.

Clinical Translation: What This Means for Prescribing

The subgroup data from SUSTAIN-6 support a simple message: semaglutide's cardiovascular benefit is not restricted to a narrow demographic. The signal was consistent across age brackets, sex, BMI strata, HbA1c levels, and renal function categories. The point estimates slightly favor older patients, those with obesity, those with worse glycemic control, and those with established CVD, but interaction tests found no statistically credible heterogeneity.

For the prescriber making a GLP-1 RA selection in a patient with T2D and cardiovascular disease, SUSTAIN-6 subgroup data do not argue against semaglutide in any tested population. The main prescribing-relevant gap is primary prevention: the trial cannot confirm MACE benefit in patients without established CVD, a limitation reflected in the approved indication.

Frequently asked questions

›

References

Marso SP, Bain SC, Consoli A, et al. Semaglutide and cardiovascular outcomes in patients with type 2 diabetes. N Engl J Med. 2016;375(19):1834-1844. PubMed
U.S. Food and Drug Administration. Ozempic (semaglutide) prescribing information. Revised 2020. FDA Label
American Diabetes Association. Standards of Medical Care in Diabetes, 2019. Diabetes Care. 2019;42(Suppl 1). PubMed
Lincoff AM, Brown-Frandsen K, Colhoun HM, et al. Semaglutide and cardiovascular outcomes in obesity without diabetes (SELECT). N Engl J Med. 2023;389(24):2221-2232. PubMed
Perkovic V, Tuttle KR, Rossing P, et al. Effects of semaglutide on chronic kidney disease in patients with type 2 diabetes (FLOW). N Engl J Med. 2024;391(2):109-121. PubMed